What Drives Cryptocurrency Returns? A Sparse Statistical Jump Model Approach

We consider the statistical sparse jump model, a recently developed, robust and interpretable regime switching model, to identify features that drive the return dynamics of the largest cryptocurrencies. The approach simultaneously performs feature selection, parameter estimation, and state classification. Our large number of candidate features comprises cryptocurrency, sentiment, and financial market-based time series that previously have been identified in the emerging literature as influencing cryptocurrency returns, as well as new ones. Our empirical study indicates that a three-state model offers the most accurate description of the cryptocurrency returns dynamics. These states have straightforward market-based interpretations as they correspond to bull, neutral, and bear market regimes, respectively. Using the data-driven feature selection methodology, we are able to determine which features are important and which ones are not. Our findings reveal that, among the set of candidate features, the first moments of returns, features that represent trends and reversal signals, market activity, and public
attention are key drivers of crypto market dynamics.